Distributional Effects of Individual Income Tax Expenditures: An Update
Eric Toder and Daniel Baneman
Urban-Brookings Tax Policy Center
February 2, 2012
ABSTRACT
Tax expenditures on average raise after-tax incomes more for upper-income than for lowerincome taxpayers. As a share of income, special rates for capital gains and dividends and
itemized deductions provide the largest benefits for taxpayers in the top 1 percent of the income
distribution, exemptions and exclusions benefit taxpayers in upper middle-income groups the
most, and refundable credits provide the largest benefits to those in the bottom two quintiles of
the distribution. Interactions among provisions make the revenue cost of all tax expenditures
about 10 percent larger than the sum of the costs of the separate provisions.
______________________________________________________________________________
The views in this paper are those of the authors alone and do not represent the views of the
Urban Institute, its board, or its funders.
The paper was presented at the National Tax Association Meetings, New Orleans, Louisiana,
November 18, 2011. An edited version of the paper will appear in an upcoming edition of the
conference proceedings.
Introduction
This paper updates estimates of the distributional effects of tax expenditures in the individual
income tax previously reported in Burman, Toder, and Geissler (2008). Compared to that
previous paper, our estimates include a more comprehensive group of individual income tax
expenditure provisions.
The paper discusses issues in measuring and interpreting distributional effects of tax
expenditures. Subject to these caveats, we then present estimates of the distributional effects of
all tax expenditures estimated simultaneously and of separate categories of tax expenditures:
exclusions from income, special rates, itemized deductions, ―above-the-line‖ deductions,
nonrefundable credits, and refundable credits. We then display information on how interactions
among provisions affect the total cost of all tax expenditures.
Issues in Measuring the Distributional Effects of Tax Expenditures
The overall distributional burden of federal taxes depends on all provisions of the tax law—the
choice of a conceptual tax base, the definition of the taxpaying unit, the tax rate schedule, and tax
preferences. The Tax Policy Center and the Congressional Budget Office (2010) have estimated
that the current federal tax system is moderately progressive, taking account of all provisions and
tax sources (individual and corporate income taxes, payroll taxes, estate and gift taxes, and
federal excise taxes).1 This moderately progressive system results from a combination of highly
progressive tax sources (the individual income tax, the corporate income tax, and estate and gift
taxes) and regressive tax sources (payroll and most excise taxes). Beyond this, the distributional
effect of the entire fiscal system also depends on who benefits from federal outlays, but
estimating this distribution is challenging because of the need to assign to income groups
benefits from public goods, such as defense spending, medical research, and environmental
protection. (For an example of a study that estimates the overall distribution of government fiscal
policies, see Chamberlain and Prante 2007.)
To estimate the distributional effect of tax expenditures in the federal income tax, an analyst
needs to divide the tax system into two sets of provisions:
1. Those provisions that are part of the ―normal‖ or baseline tax system, and
2. Those that are labeled tax expenditures because they are ―special‖ provisions, or
exceptions to the general rules, that benefit selected taxpayers or encourage selected
activities.
For the estimates in this paper, we use the Office of Management and Budget (OMB) definition
of tax expenditures (Office of Management and Budget 2011). Therefore, for this paper, we do
1
TPC currently does not include excise taxes and customs duties in its estimates of the distribution of the burden of
federal taxes, but excise taxes and customs duties compose a small share of total federal tax receipts.
2
not address the question of which departures from income measurement should be viewed as a
general feature of the federal income tax and which provisions should be viewed as a disguised
spending program administered through the income tax (Marron and Toder 2011). Instead, we
simply estimate the effects of the subset of tax provisions that OMB, and in most cases the Joint
Committee on Taxation (JCT), calls tax expenditures.
Nonetheless, it is important to recall that progressivity reflects all provisions of federal taxes. For
example, the same steeply graduated rates that make the distribution of the federal income tax
very progressive also make tax exemptions conveyed in the form of a reduction in taxable
income (exemptions, deductions, deferrals of income recognition) regressive because with rising
marginal rates upper-income taxpayers receive the biggest benefits from these provisions.
The estimates in this paper include only the effects of losses in revenue from individual income
taxes, although exclusion of some forms of income from the individual income tax base also
reduces payroll tax liability. One reason for omitting the distributional effects of provisions that
affect payroll tax liability is that they are difficult to interpret. Provisions that reduce taxable
earnings reduce both payroll tax liability and incremental future Social Security retirement and
disability benefits associated with additional taxable earnings. But incremental Social Security
benefits associated with any given increase in a single year’s taxable earnings differ greatly
among individuals, depending on their lifetime earnings, years of covered earnings, marital
status, and earnings of their current, former, or future spouses. This makes it difficult to assess
the net effect on any individual of preferences that reduce their payroll tax liability.
There are a number of reasons to interpret distributional estimates of tax expenditures with
caution. Toder, Harris, and Lim (2011) discuss five issues in interpreting the estimates. First, as
noted above, the choice of some components of the baseline tax system is arbitrary, so that the
definition and size of tax expenditures can vary depending on how one defines the baseline tax
law. Second, the revenue loss from a tax expenditure provision in any single year does not
accurately measure its benefit to taxpayers for those tax expenditures that alter the timing of tax
payments. Third, the tax expenditure estimate for a provision may overstate the burden on
taxpayers from eliminating the provision if they can change their behavior to escape a portion of
the additional tax. Fourth, the economic incidence of tax expenditures may differ from the
incidence assumed in the estimates. Finally, the net effect of eliminating tax expenditures
depends on how the revenue gain is used. The last two issues are worth some additional
discussion.
Incidence Assumptions
The Tax Policy Center and the government agencies that perform distributional estimates (the
Office of Tax Policy at the U.S. Treasury Department, the Joint Committee on Taxation, and the
Congressional Budget Office) all assume that individuals bear the burden of individual income
taxes in proportion to their tax liability. This means that pretax earnings and market prices of
3
different consumption goods are assumed to be unaffected by changes in individual income tax
provisions.
For some provisions, however, this assumption might be questionable. For example, estimators
assume that eliminating the exemption of interest on tax-exempt bonds would raise tax burdens
only on current holders of tax-exempt securities. But tax exemption in current law makes the
pretax interest rates on exempt securities lower than the yields on taxable securities of
comparable maturity and risk. If tax exemption were eliminated, relative pretax yields on taxexempt securities would rise and yields on taxable securities would fall, shifting some of the
burden of the tax change from current holders of tax-exempt bonds to all capital income
recipients. There would also be changes in relative prices facing users of capital services—costs
to state and local borrowers would increase and costs to private sector borrowers would decline.
But these user side effects of relative price changes are typically ignored in estimates of the
distributional effects of individual income tax changes.2
A more complex incidence issue is how to treat exemption of employee fringe benefits that must
be provided uniformly to all or a given group of employees. TPC follows the assumption
estimators typically use that fringe benefits substitute for cash wages on a dollar-for-dollar basis
so that, for example, the benefit employees receive from tax-free employer-provided benefits is
equal to the tax that would otherwise be paid if those benefits were taxed as wages. If, however,
the tax-free status of fringe benefits causes employers to change the distribution of pretax
compensation among workers (because, for example, high tax-bracket workers value tax-free
fringe benefits more than low tax-bracket workers), then estimators may overstate the relative net
benefits that high-income workers receive from employer-provided fringe benefits. For example,
Smith and Toder (2011) report evidence that employers reduce wages of high-income employees
more per dollar of employer contributions to 401(k) plans than they reduce wages of low-income
employees.
Uses of Revenue
Whether tax expenditures make the tax law more or less progressive depends on how the revenue
from eliminating tax expenditures might be used. The additional revenue from eliminating tax
breaks could be used to reduce marginal tax rates—either uniformly or in some other pattern—or
could be used to fund new spending programs that benefit all taxpayers uniformly or benefit lowincome households relatively more. Without knowing how an increase in revenue would be spent
(or a reduction financed), one cannot definitively estimate the distributional effects of any policy
change that is not budget neutral.
2
While user side effects on particular economic sectors are significant, their effects on the distribution of tax
burdens across income groups may be secondary. But we can’t assert that for certain.
4
TPC presents as its main distributional metric the percentage change in after-tax income for each
income group that a tax law change would produce. This measure would correctly identify as net
losers from eliminating tax expenditures, or net winners from existing tax expenditures, those
with a larger than average decline in after-tax income if the additional revenue from eliminating
tax expenditures were used to fund tax cuts or benefits that were proportional to after-tax income
for all taxpayers. An alternative metric is the percentage change in tax liability associated with a
particular proposal. This metric would correctly identify as net winners from existing tax
expenditures those with a larger than average percentage increase in taxes paid if the revenue
raised were used to finance proportional across-the-board tax cuts for all taxpayers. A third
possible metric is the absolute increase in tax liability associated with eliminating tax
expenditures. This metric would correctly identify as net winners from existing tax expenditures
those with a larger than average absolute increase in tax liability if the revenue raised were used
to finance spending programs that provide the same absolute dollar benefit to all taxpayers.
With these qualifications and cautions duly noted, we forge ahead and present our latest
estimates of the distributional effects of individual income tax expenditures.
Estimates of Distributional Effects of Individual Income Tax Expenditures
Taxpayers at all income levels receive benefits from tax expenditures, but high-income taxpayers
receive larger benefits on average as a share of after-tax income from tax expenditures than do
low-income taxpayers. The distribution of benefits, however, differs widely across different
forms of tax benefits.
Overall Distributional Effect of Eliminating Tax Expenditures
Overall, eliminating tax expenditures would reduce after-tax income by 12.3 percent (table 1) in
tax year 2011, but reduce after-tax income by less than the average amount for all income groups
in the bottom 90 percent of the population and reduce after-tax income by more than the average
amount for all groups in the top 10 percent. Eliminating tax expenditures would reduce after-tax
income by 19.8 percent for taxpayers in the top 1 percent of the distribution, compared with only
7.5 percent in the bottom quintile. The percentage decline in after-tax income rises as income
rises, except for the second quintile, where taxpayers would see a larger tax increase as a share of
after-tax income than those in the third and fourth quintiles in the distribution. These taxpayers
currently receive substantial benefits as a share of their income from the child credit and earned
income credit.
Similarly, taxpayers in the top income groups would receive a larger share of tax increase from
eliminating tax expenditures than their shares of pretax income (compare columns 2 and 3 of
table 1). Taxpayers in the top 1 percent receive about 17 percent of income, but would bear 24
percent of the cost of eliminating tax expenditures. Taxpayers in the top fifth of the distribution
5
receive about 55 percent of income, but would pay 66 percent of the increase in taxes. This
means that a reform that eliminated tax expenditures and gave the money back to taxpayers as a
tax cut (or grant) equal to a constant proportion of income would make the tax system more
progressive.
The picture looks a little bit different, however, if one compares shares of tax increases from
eliminating tax expenditures with shares of taxes paid under current law (compare columns 2 and
4 of table 1). High-income taxpayers would bear a slightly lower share of the cost of eliminating
tax expenditures than the share of taxes they currently pay (24 percent of the tax increase versus
26 percent of current taxes paid for the top 1 percent; 66 percent of the tax increase versus 70
percent of current taxes paid for the top quintile). Taxpayers in the fourth quintile would also pay
a lower share of the tax increase than the share of taxes they currently pay. In contrast, taxpayers
in the bottom two quintiles would pay about 12 percent of the increase in taxes but currently pay
less than 3 percent of federal income taxes. This means that a tax reform that removed all tax
expenditures and gave the revenue back in the form of an equal proportional marginal rate cut for
all taxpayers would make the tax system less progressive.3
Distribution by Categories
The distributional effect among income groups of eliminating tax expenditures varies widely
among tax expenditure categories, with the categories based on the form in which the subsidy is
conveyed. (Appendix 1 lists the provisions in each category.) Relative to the population as
whole, high-income taxpayers would lose the most from eliminating special rates for capital
gains and dividends, but also bear disproportionate costs as a share of after-tax income from
eliminating exclusions and itemized deductions (table 2). In contrast, low-income taxpayers
would lose the most from elimination of refundable credits.
Exclusions are those provisions that exempt some income from tax and represent the largest
category of tax expenditures. The largest five exemptions are those for (1) employer-sponsored
health insurance benefits and health benefits under section 125 cafeteria plans, (2) income
accrued within qualified retirement plans, (3) net imputed rental income on owner-occupied
homes, (4) capital gains transferred at death (step-up in basis), and (5) capital gains on home
sales. The benefits of exclusions are widely distributed, reflecting different effects of different
provisions, but on average upper middle-income taxpayers (those in the 90th to 99th percentile of
the distribution) receive the largest proportional benefits from exclusions. These taxpayers
benefit most from the tax preferences for retirement saving. In contrast, taxpayers in the middle
3
In contrast, replacing tax expenditures with an equal percentage point rate cut for all taxpayers (and a subsidy of
the same percentage of income for those with no tax liability) would make the tax system somewhat more
progressive, and replacing them with an equal per capita refundable credit for all taxpayers would make the tax
system much more progressive. For illustrations of the effects of replacing tax expenditures with alternative forms of
tax cuts, see Toder, Harris, and Lim (2011).
6
of the distribution benefit most in proportion to their income from the exclusion of employersponsored health insurance benefits. The very highest income taxpayers receive less than
proportional benefits from the health insurance exclusion (because premiums do not rise much at
the highest incomes) and from qualified retirement plans (because contributions to defined
contribution plans and benefits from defined benefit plans are capped), but gain a major share of
the benefits from the exclusion of capital gains transferred at death.
Net long-term capital gains and qualified dividends are taxed at rates of 0 and 15 percent,
compared to ordinary income rates that range from 10 to 35 percent. The very highest income
taxpayers receive a large share of capital gains and dividends and gain the most from the rate
differential. As a result, eliminating these preferences would reduce after-tax income by an
average of 4.5 percent in the top 1 percent of the distribution but by less than 1 percent on
average for taxpayers in all other income groups, including those in the 95th–99th percentiles.
Itemized deductions provide the largest percentage benefit as a share of after-tax income to
taxpayers in the top 1 percent of the income distribution, largely due to the concentration of large
charitable donations in this group. Taxpayers in the 80th–99th percentiles are the largest
beneficiaries of the mortgage interest deduction (Toder et al. 2010). Taxpayers in the lowestincome groups benefit little from itemized deductions because most of them claim the standard
deduction and those who do itemize are in lower tax brackets and receive relatively less benefit
from a deduction than those in high tax brackets.
Above-the-line deductions and nonrefundable credits account for a relatively small share of all
tax expenditures. The largest above-the-line deductions that are tax expenditures are the
deduction for medical insurance premiums for the self-employed and the additional standard
deduction for the blind and elderly.4 The benefit from above-the-line deductions is distributed
fairly evenly across income groups, except for the bottom quintile. In that group, many have no
tax liability and therefore do not benefit from additional deductions. The largest nonrefundable
credits are the child and dependent care credit, the savers’ credit, the general business credit, and
the lifetime learning credit. Some of these credits either phase out or phase down with income,
and capped credits in general decline as a share of income as income rises. But nonrefundable
credits cannot be used by taxpayers who have no tax liability. As a result, the largest proportional
benefit from these credits goes to taxpayers in the second and third quintiles of the income
distribution. Many taxpayers in the bottom quintile cannot use them and taxpayers in the highest
income quintiles receive reduced benefits from some credits and no benefits from others.
Refundable credits, in contrast, provide the largest proportional benefits as a share of after-tax
income to taxpayers in the bottom two quintiles of the income distribution. The three largest
4
We count the deductions for contributions to retirement plans in the exclusions line because the largest net
component of the retirement preference is the exemption of income accrued within qualified plans.
7
refundable credits in tax year 2011 are the earned income tax credit, the child tax credit, and the
American opportunity tax credit. The benefits we show include both the refundable portion
(counted as spending in budgetary presentations) and the portion that reduces tax liability to zero
(counted as a tax reduction in budgetary presentations). Middle-income taxpayers do receive
substantial benefits from the child credit and the American opportunity tax credit, but the benefit
as a share of after-tax income falls as income rises because the credit per child is a fixed amount,
the American opportunity tax credit is capped, and both credits phase out at higher income
levels.
The final category—miscellaneous tax expenditures—consists mostly of small provisions that
TPC estimated off-model by using OMB figures for the size of the tax expenditure and
distributing the benefit in proportion to the income source or sources affected (capital gains,
interest income or earnings) multiplied by the applicable marginal tax rate on the income source
or sources. The provisions in this category are shown in appendix 1. Eliminating them would
generally raise taxes by a larger share of income for higher-income taxpayers than for lowerincome taxpayers.
In summary, the distribution of benefits from tax expenditures varies substantially among types
of provisions (table 3). Taxpayers in the top 1 percent of the distribution pay 26 percent of
federal taxes but would pay 75 percent of the additional taxes from elimination of the special
rates for capital gains and dividends. They would pay about 26 percent of the increased taxes
from elimination of itemized deductions, about the same as the share of the taxes they currently
pay. For all other forms of tax expenditures, they would pay a lower share of the increased taxes
from elimination of the provisions than the share of federal taxes they currently pay. Other
taxpayers in the top quintile of the distribution (80th–99th percentiles) pay 44 percent of federal
taxes, but would pay 55 percent of the higher taxes from elimination of itemized deductions and
51 percent of the higher taxes from elimination of exclusions. Their share of the higher taxes
from elimination of above-the-line deductions and other miscellaneous provisions is about the
same as the share of taxes they currently pay. But they would pay relatively less additional tax,
in relation to the taxes they currently pay, from the elimination of nonrefundable credits, special
rates for capital gains and dividends, and refundable credits. Taxpayers in the third and fourth
quintiles combined would pay relatively more tax from elimination of credits (both
nonrefundable and refundable) and above-the-line deductions than the taxes they currently pay,
about the same as their share of current taxes from elimination of exclusions, and a smaller share
relative to their current share of taxes from elimination of other provisions. Taxpayers in the
bottom two quintiles pay a very small share of federal taxes. They would lose the most from
elimination of refundable credits but would also bear more than their current share of taxes of the
costs of elimination of nonrefundable credits, exclusions, and above-the-line deductions. But the
lowest-income taxpayers would pay virtually none of the increased taxes from the elimination of
itemized deductions and the special rates for capital gains and dividends.
8
Effects of Interactions
The Tax Policy Center (TPC) estimates presented in this paper show the effects of eliminating all
individual tax expenditures simultaneously. OMB and JCT, however, do not report this
information. What they do report is an estimate of the revenue loss from each separate tax
expenditure provision assuming all other provisions of the income tax law remain in effect. But
if some tax expenditures were eliminated, the revenue effect of eliminating other tax
expenditures would differ from those reported by OMB and JCT. The separate tax expenditure
estimates cannot be added up to obtain a total cost of tax expenditures and OMB and JCT
therefore do not add them up.
Nonetheless, many analysts are interested in the total cost of tax expenditures and how those
costs have changed over time (for example, see Rogers and Toder 2011). These analysts do add
up tax expenditures, while acknowledging that the totals are inexact because they ignore
interactions among provisions. Burman, Toder, and Geissler (2008) used the TPC model to
compare estimates of the total cost (including interactions) of tax expenditure provisions to the
sum of the individual costs (ignoring interactions) of all provisions under alternative assumptions
about what the alternative minimum tax provisions for 2007 would turn out to be and found that
interactions raised the cost of the tax expenditures they included in their paper by between 5 and
8 percent.5
For tax year 2011, TPC now estimates that interactions raised the total revenue gain from
eliminating all individual income tax expenditures by 9.6 percent, compared with the sum of the
gains from eliminating all provisions separately (table 4). The effects of interactions among
provisions differ among types of provisions. For all provisions except itemized deductions,
interactions raise or leave unchanged the revenue gain from eliminating tax expenditures. The
largest effect among these provisions is for exclusions; interactions raise the revenue gain from
eliminating exclusions by 4.7 percent because eliminating an exclusion of some income pushes
taxpayers into higher marginal rate brackets, thereby increasing the gain from eliminating other
exclusions. But interactions have a dramatically different effect for itemized deductions,
reducing their total cost by 27.3 percent. This negative interaction occurs because eliminating
any one itemized deduction pushes more taxpayers onto the standard deduction, thereby reducing
the revenue gain from eliminating other itemized deductions.
5
At the time the estimates for the Burman, Toder, and Geissler (2008) paper were made, the temporary increase in
exemptions under the individual alternative minimum tax (the AMT patch) had expired for tax year 2007, but
Congress was considering a last-minute fix before taxpayers filed their 2007 income tax returns (which they
eventually enacted). The paper’s estimates were made using two extreme assumptions: (1) current law AMT,
meaning the patch was not extended, and (2) no AMT. The simulations for this paper assume that the AMT is
eliminated as part of the simulation that eliminates all tax expenditures, but the estimates of specific provisions and
groups of provisions assume the 2011 patched AMT remains in effect. Also, Burman, Toder, and Geissler did not
include all the tax expenditure provisions that we estimate in this paper.
9
Across all categories, the sum of the individual revenue gains from eliminating tax expenditures,
without considering interactions, is $985.9 billion. After including interactions within categories,
the total revenue gain becomes $955.6 billion, about 3.2 percent less than the sum of the separate
provisions. However, incorporating interactions across categories increases the total revenue
gain from eliminating tax expenditures to $1,080.6 billion, about 9.6 percent more than the sum
of the separate provisions. This is driven primarily by the interaction of special capital gains rates
with the capital gains exclusion on home sales and step-up in basis of capital gains at death, as
the elimination of the special capital gains rates significantly increases the value of these
exclusions.
Conclusions
This paper presents new estimates of the distributional burden of tax expenditures in the
individual income tax. Taxpayers in all income groups benefit from tax expenditures, but we find
that, on average, tax expenditures raise after-tax incomes more for upper-income taxpayers than
for lower-income taxpayers. Taxpayers in the second quintile of the distribution, however,
receive larger proportional benefits from tax expenditures than middle-income taxpayers in the
third and fourth quintiles of the distribution.
The benefits of tax expenditures vary substantially by type of provision. Special rates for capital
gains and dividends provide much larger proportional benefits for taxpayers in the top 1 percent
of the distribution than for others. Itemized deductions also provide the largest proportional
benefit for taxpayers in the top 1 percent of the distribution, but provide substantial benefits as
well for other taxpayers in the top quintile. Exclusions and exemptions benefit taxpayers in the
upper middle-income groups the most, but also provide substantial benefits to the taxpayers in
the middle and at the very top of the income distribution. Above-the-line deductions and
nonrefundable credits are relatively small, but their benefits are widely spread and are larger as a
share of after-tax income for those in the middle of the income distribution. Finally, refundable
credits—principally the earned income tax credit and the child tax credit—provided the largest
benefits as a share of income to those in the bottom two quintiles of the income distribution.
In spite of the regressive overall distribution of tax expenditures, it is worth emphasizing that the
overall federal tax system is moderately progressive. One reason that many tax expenditures
provide relatively larger benefits to those in upper income groups is that taxpayers in these
groups face higher marginal tax rates under the ―baseline‖ tax system against which the tax
expenditure provisions are measured.
Finally, we find that the revenue gain from eliminating all individual income tax expenditures
simultaneously is almost 10 percent larger than the sum of the gains from eliminating separate
tax expenditure provisions. The aggregate figures reflect large differences within tax expenditure
categories. While the gain from eliminating all tax expenditures exceeds the sum of gains from
individual provisions within most tax expenditure categories, eliminating all itemized deductions
10
raises less than 75 percent of the revenue from summing the estimates of the gains from the
separate itemized deductions. This anomalous result for itemized deductions occurs because each
time an itemized deduction is eliminated, more taxpayers move onto the standard deduction,
thereby reducing the gain from eliminating remaining deductions. In spite of this, however,
combining all the tax expenditures raises the estimated gain above the sum of the gains from all
the separate items, largely because eliminating some tax expenditures pushes taxpayers into
higher rate brackets for both ordinary income and capital gains, making the pickup from
eliminating others larger. This means that estimates of the total cost of tax expenditures derived
by adding up individual provisions understate their total budgetary cost.
*Eric Toder is codirector, Urban-Brookings Tax Policy. Daniel Baneman is a research assistant
at the Urban Institute, Tax Policy Center.
11
References
Burman, Leonard, Eric Toder, and Christopher Geissler (2008). ―How Big Are Total Individual
Income Tax Expenditures and Who Benefits from Them?‖ Urban-Brookings Tax Policy Center.
Discussion Paper No. 31. December.
Chamberlain, Andrew, and Gerald Prante (2007). ―Who Pays Taxes and Who Receives
Government Spending? An Analysis of Federal, State, and Local Tax and Spending
Distributions, 1991–2004.‖ Tax Foundation Working Paper 1. March.
Congressional Budget Office (2010). ―Average Federal Taxes by Income Group.‖
http://www.cbo.gov/publications/collections/collections.cfm?collect=13.
Marron, Donald, and Eric Toder (2011). ―How Big Is the Federal Government?‖ Presented at
National Tax Association Meetings. New Orleans, Louisiana. November.
Office of Management and Budget (2011). The Fiscal Year 2012 Budget of the United States
Government. Analytical Perspectives.
Rogers, Allison, and Eric Toder (2011). ―Trends in Tax Expenditures, 1985–2016.‖ UrbanBrookings Tax Policy Center. September 16.
Smith, Karen E., and Eric Toder (2011). ―Do Low-Income Workers Benefit from 401(k) Plans?‖
Center for Retirement Research at Boston College. Working Paper 2011-14. September.
Toder, Eric, Benjamin Harris, and Katherine Lim (2011). ―Distributional Effects of Tax
Expenditures in the United States.‖ In Lisa Philipps, Neil Brooks, and Jinyan Li, editors, Tax
Expenditures: State of the Art. Toronto, ON: Canadian Tax Foundation. 4:2–4:35.
Toder, Eric, Margery Austin Turner, Katherine Lim, and Lisa Getsinger (2010). ―Reforming the
Mortgage Interest Deduction.‖ Urban Institute, Urban-Brookings Tax Policy Center, and What
Works Collaborative Working Paper. April.
12
Table 1. Effect of Eliminating Nonbusiness Individual Income Tax Expenditures, 2011
Cash Income
Percentile
Lowest quintile
2nd quintile
3rd quintile
4th quintile
80–90th
percentiles
90–95th
percentiles
95–99th
percentiles
Top 1 percent
Total
Percent Change
in After-Tax
Income
-7.5 percent
-9.8 percent
-8.1 percent
-8.4 percent
-12.2 percent
Share of Tax
Change
Share of Income,
Current Law
2.8 percent
7.8 percent
9.5 percent
13.8 percent
14.1 percent
3.8 percent
8.5 percent
13.5 percent
19.9 percent
14.3 percent
Share of Tax
Liability, Current
Law
0.2 percent
2.7 percent
9.3 percent
18.2 percent
15.1 percent
-15.0 percent
11.5 percent
9.7 percent
11.3 percent
-15.9 percent
16.7 percent
13.8 percent
17.5 percent
-19.8 percent
-12.3 percent
23.9 percent
100 percent
16.8 percent
100 percent
25.6 percent
100 percent
13
Table 2. Percentage Changes in After-Tax Income from Eliminating Various Categories of
Individual Income Tax Expenditures
Cash income
percentile
Exclusions Capital
Itemized
gains and deductions
dividends
-0.9
*
*
Abovethe-line
deductions
*
NonRefundable Other All
refundable
credits
credit
*
-6.0
-0.1
-7.5
-4.2
*
-0.1
-0.1
-0.2
-5.5
-0.3
-9.8
3rd quintile
-4.7
-0.1
-0.4
-0.1
-0.2
-2.0
-0.4
-8.1
4th quintile
-4.4
-0.1
-1.2
-0.1
-0.1
-1.0
-0.5
-8.4
80–90
-7.0
-0.2
-2.2
-0.1
-0.1
-0.6
-0.7
-12.2
90–95
-8.9
-0.4
-2.8
-0.1
-0.1
-0.2
-0.8
-15.0
95–99
-9.4
-0.9
-2.6
-0.1
*
*
-1.1
-15.9
Top 1
-6.4
-4.5
-3.0
-0.1
-0.1
*
-1.6
-19.8
ALL
-6.0
-0.9
-1.7
-0.1
-0.1
-1.4
-0.8
-12.3
Lowest
quintile
2nd quintile
*=absolute value less than 0.005 percent
Note: Totals for a row may not add up to the sum for the row because of interactions among
provisions.
14
Table 3. Distribution of Benefits of Various Categories of Individual Income Tax
Expenditures
Cash
income
percentile
Lowest
quintile
2nd
quintile
3rd
quintile
4th
quintile
80–90
90–95
95–99
Top 1
Exclusions Capital
Itemized
gains and deductions
dividends
0.7%
*
*
Abovethe-line
deductions
0.4%
NonRefundable Other
refundable
credits
credit
1.2%
19.7%
0.4%
0.2%
6.8%
0.2%
0.7%
5.7%
16.4%
38.6%
3.4%
2.7%
11.2%
0.9%
3.8%
17.5%
26.4%
20.4%
8.0%
9.3%
14.8%
2.8%
14.2%
24.3%
25.8%
14.1%
14.0% 18.2%
16.5%
13.9%
20.3%
15.9%
3.5%
4.2%
13.3%
75.1%
18.6%
15.9%
20.4%
26.4%
18.1%
9.4%
16.3%
8.3%
12.5%
5.7%
3.8%
8.3%
5.7%
1.0%
0.2%
*
12.4%
10.2%
19.4%
32.2%
*absolute value less than 0.005 percent
Note: Totals for a row may not add up to the sum for the row because of interactions among
provisions.
15
Taxes
Paid
15.1%
11.3%
17.5%
25.6%
Table 4. Effects of Interactions on Tax Expenditure Estimates (in billions of dollars)
Type of Provision
Exclusions
Above-the-line
deductions
Special rates for
capital gains and
dividends
Itemized deductions
Nonrefundable credits
Refundable credits
Miscellaneous
provisions
Sum of all categories
Total, all provisions
Total Cost without
Interactions
502.3
8.8
Total Cost with
Interactions
525.7
8.8
Percentage Change
Due to Interactions
4.7%
0.2%
76.4
77.7
1.7%.
202.2
7.9
121.9
66.4
147.0
8.0
122.0
66.4
-27.3%
1.2%
0.1%
n.a.
985.9
985.9
955.6
1080.6
-3.2%
+9.6%
16
Appendix
Tax Expenditures by Category Included in TPC Estimates of Total Cost of Tax
Expenditures
Exclusions










Exclusion of net imputed rental income
Exclusion of interest on tax-exempt bonds
Exclusion of Social Security benefits in excess of 15 percent
Exclusion of workers’ compensation benefits
Exclusion of inside buildup on life insurance and annuities
Exclusion of income earned abroad by U.S. citizens
Exclusion of employer-sponsored health insurance benefits and benefits under Section
125 cafeteria plans
Tax-deferred IRAs and exclusion of inside buildup in defined benefit and defined
contribution retirement plans
Step-up in basis of capital gains at death
Capital gains exclusion on home sales
Above-the-Line Deductions






Student loan interest deduction
Self-employed medical insurance premium deduction
Health savings account deduction
Deduction for certain education expenses
Educator expense deduction
Additional standard deduction for the elderly and blind
Capital Gains and Dividends


Preferential tax rates on long-term capital gains
Preferential tax rates on qualified dividends
Itemized Deductions




Deductibility of mortgage interest on owner-occupied homes
Deductibility of state and local real estate taxes and income or sales taxes
Deductibility of charitable contributions
Deductibility of medical expenses
17
Nonrefundable Credits





Lifetime learning credit
Child and dependent care credit
Saver’s credit
Credit for the elderly or disabled
General business credit
Refundable Credits



Child tax credit
Earned income credit
American opportunity tax credit
Other Exclusions and Miscellaneous Tax Expenditures










Exclusion of veterans’ death benefits and disability compensation
Exclusion of premiums on group term life insurance
Exclusion of scholarship and fellowship income
Exclusion of unreimbursed employee parking expenses
Exclusion of capital gains on small corporate stock
Deferral of income from installment sales
Exception from passive loss rules for $25,000 of rental loss
Education individual retirement accounts
State prepaid tuition plans
Other exclusions and miscellaneous tax expenditures
18